adversarial-beam

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly requires composing a self-contained reviewer prompt that includes "stack facts (from mix.exs and config/)" and also requires evidence as a file:line or a quote, so if config files contain API keys or secrets the LLM would be asked to include those secret values verbatim in the prompt or evidence output.